OBO challenge 2023: go green in Uberon OBO dashboard

[paolaroncaglia]

Similarly to https://github.com/obophenotype/cell-ontology/issues/1262, placeholder ticket inspired by @matentzn 's talk at ICBO 2021. Here's the Uberon OBO dashboard report.

[matentzn]

Thats the spirit! For technical reasons, the ROBOT report on the dashboard is truncated, but if you want the full list as a table, I will generate it for you!

[dosumis]

Full would be useful. I think it's not as bad as it looks at first glance. e.g. - some obvious signs of pipeline/declaration issues.

[paolaroncaglia]

Self-assigning for shepherding purposes. Not urgent, but to start with,

• [x] @matentzn could you please provide the full report.

We can then better assess what needs to be done, and think about assigning tasks. Some, like adding missing definitions, could be labelled Easy First Issue and kept for training purposes if there are plans for new curators starting soon-ish (or for anyone wishing to contribute).

[matentzn]

https://github.com/obophenotype/uberon/blob/master/src/ontology/reports/uberon-edit.obo-obo-report.tsv

You can refresh with:

cd src/ontology
sh run.sh make IMP=false PAT=false robot_reports

but it will also be refreshed with every release.

[paolaroncaglia]

@matentzn Thanks for the speedy reply! May I please ask to (no rush)

• [ ] Add a column with IDs (useful to sort and filter out, e.g. some missing definitions are for non-Uberon terms)
• [ ] Look into the INFO missing_superclass message. Terms I created are in there because they have an equivalent axiom and (as per David's instructions) I did not add the redundant superclass, e.g. the last one in the list. I checked a couple of terms created by others and they're the same, and there's a ton of them. (David suggested - signs of pipeline/declaration issues.) Should the missing_superclass check be changed to look for real cases, or excluded from the ROBOT dashboard report, or the check name changed to something less alarming? :-)

Then we can better focus on the error/warnings. Thanks.

[paolaroncaglia]

P.S. @matentzn, I'll add you as assignee as you've worked on this.

[matentzn]

I would only care about ERROR for now.. INFO ones leave them for last.

[paolaroncaglia]

Update re. Uberon OBO dashboard for the latest version of Uberon (released Feb 21st 2022):

1. 130 errors due to duplicate_definition
2. 1145 warnings due to duplicate_scoped_synonym
3. 2 missing_subset_declaration (not found in GitHub report)
4. 743 warnings due to duplicate_label_synonym
5. 2863 warnings due to missing_definition
6. 187 info about lowercase_definition
7. 2818 info about missing_superclass

Entries in 1. are classified as warnings, not errors, in the report generated in GitHub as part of the release, and have a dedicated ticket.

Entries in 2. also have a dedicated ticket .

Entries in 4. (I've spot-checked a few) are likely due to import of xref labels (e.g. from FMA) and seem to all affect same-same term (i.e. a term has an exact synonym that's identical to its primary label). These particular cases should probably not give rise to warnings; it's far more serious when an exact synonym of term A is identical to primary label of term B, but there are checks in place for that.

Entries in 5. will require a significant effort to tackle. Note they are not all Uberon terms - but mostly. Ideas: Prioritise terms that are not in the UBERON:2X, 3X, 4X, 6X? See if terms have logical axioms that could help generate automated textual definitions, similarly to https://github.com/obophenotype/cell-ontology/issues/1262#issuecomment-1046759234?

[paolaroncaglia]

@matentzn FYI I'll remove my assignment and add to the tech board. I hope the outline above helps ;-)

[shawntanzk]

@shawntanzk to look into non-RO relations http://dashboard.obofoundry.org/dashboard/uberon/fp7.html

eg http://purl.obolibrary.org/obo/uberon/core#synapsed_by should be a RO term - look into why it is there

Related to https://github.com/obophenotype/uberon/issues/1823

[github-actions[bot]]

This issue has not seen any activity in the past 6 months; it will be closed automatically one year from now if no action is taken.

[anitacaron]

Currently, we have 129 errors related to the text definition. 126 of them are related to the definition being a simple dot. This will be fixed with the new ODK release.

Another error type is using UBERON namespace for two RO properties.

• [x] https://github.com/obophenotype/uberon/issues/2811

[github-actions[bot]]

This issue has not seen any activity in the past 6 months; it will be closed automatically one year from now if no action is taken.

[anitacaron]

Update today. The most urgent one is the duplicate definition error

[anitacaron]

Now that PR #2964 has been merged, Uberon won't have any errors in the next dashboard release.

[anitacaron]

Uberon is green in the last dashboard release (2024-01-24) 🎉